Skip to content

Mistral AI · Chat

Magistral Medium

Mistral AI 🇫🇷ChatMagistral familyOpen weights. One of its sources says the weights are publicly available.

Official mistral-medium-latest Mistral model

llminfo IQ
112.6
Maker's price
$2.75 blended / 1M
Max context
262K
Listings
11

Evidence and sources

llminfo IQ 112.6, ranked among chat models. From 69 benchmarks. Plenty of data behind it.How llminfo IQ is computedCoding IQ 91.6, 95% interval 82–101. From 5 benchmark families across 4 sources. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.How Coding IQ is computedPoor value (22.3). Costs more per task than its benchmark results predict. From 15 benchmarks with measured task costs. Plenty of data behind it.How Value for money worksReleased 9 Jun 2025Maker HQ: FR (according to Mistral AI's own website)
Sources: Mistral API data, LMArena data, Models.dev data, llminfo researchPrimary data from Mistral API from LMArena: name, family, variant from Models.dev: weights availability from llminfo research: release date
Tags from its sources:premierreasoningmultilingual

Advantages and trade-offs

Compared with 10 chat models, picked mainly by popularity and IQ

Advantages

Open weights available

The weights are public. 2 of the 7 peers that say either way have open weights.

We only make a claim when there's enough data on models of the same type. Updated 22 Nov, 06:07 UTC.

Input & output

as stated by the model's providers
Text
Images
Files
Text

Capabilities

Supported operations and provider-stated features

VisionReasoningFunction callingStructured output

Listings

 Mistral AImagistral-medium-latest · latestSources: LiteLLM data, Mistral API data, Models.dev data262K$1.50cache read $0.15$7.50
 Mistral AImagistral-mediumSource: GenAI Prices dataNot published by the provider$2.00$5.00
 Eden AIInference by Mistral AImistral/magistral-medium-latest · latestSource: Models.dev data262K$1.50cache read $0.15$7.50
 Eden AIInference by Mistral AIRegions: 🇪🇺 eumistral/magistral-medium-latest@eu · latestSource: Models.dev data262K$1.65cache read $0.165$8.25
 Euroutermagistral-mediumSource: EURouter data40K$2.20$5.50
 Flowdotmagistral-medium-latest · latestSource: FlowDot data40K$2.00$5.00
 OpenRoutermagistral-medium-2506 · 2506Source: GenAI Prices dataNot published by the provider$2.00$5.00
 OpenRoutermistralai/magistral-medium-2506:thinking · thinking, 2506Source: Helicone Costs dataNot published by the provider$2.00$5.00
 Vercelmistral/magistral-mediumSource: Helicone Costs dataNot published by the provider$2.00$5.00
 Merge Gatewaymistral/magistral-medium-latest · latestSource: Models.dev data128K$2.00$5.00
 Vercel AI Gatewayvercel_ai_gateway/mistral/magistral-mediumSource: LiteLLM data128K$2.00$5.00

Benchmarks

Named effort levels run from lowest to highest. Select a result to see its source details.

Human preference

Agentic

Reasoning

Coding

Math

AIME 2024Source: LLM Stats data · Reported by the model's developer. The source relaying it did not verify it. claim
AIME 2025Source: LLM Stats data · Reported by the model's developer. The source relaying it did not verify it. claim

Safety

PhareSource: Phare

Factuality

Instruction following

Science

GPQASource: LLM Stats data · Reported by the model's developer. The source relaying it did not verify it. claim

Other

Data credits: Data by LLM Stats · Epoch AI Benchmarking Hub, CC BY 4.0 (licence) · LMArena, CC BY 4.0 (licence) · Galileo Agent Leaderboard, Apache-2.0 (licence) · Phare (Giskard), CC BY 4.0 (licence) · SnakeBench (Greg Kamradt / ARC Prize Foundation), MIT (licence)

Around the web

  • Magistral | Mistral AI
    Mistral AI © 2026 Sélectionner la langue Français English Français Italiano Highlights. Plot Magistral Magistral …Mistral AI © 2026 Sélectionner la langue Français English Français Italiano Highlights. Plot Magistral Magistral is a dual-release model focused on real-world reasoning and feedback-driven improvemen…

    mistral.ai

  • Magistral | Mistral AI
    Magistral · Mistral AI Highlights. Magistral is a dual-release model focused on real-world reasoning and …Magistral · Mistral AI Highlights. Magistral is a dual-release model focused on real-world reasoning and feedback-driven improvement. We’re releasing the model in two variants: Magistral Small — a 24…

    mistral.ai

  • Magistral-Small-2509-Q4_K_M.gguf · mistralai/Magistral-Small-2509-GGUF at dca37996ac7810e9c1ed4ab89401dafebf20d75f
    mistralai / Magistral-Small-2509-GGUF like 72 Follow Mistral AI_ 16.6k GGUF 24 languages llama.cpp License: apache-2.0 …mistralai / Magistral-Small-2509-GGUF like 72 Follow Mistral AI_ 16.6k GGUF 24 languages llama.cpp License: apache-2.0 Model card Files Files and versions xet Community 4 Use this model dca3799 Magis…

    huggingface.co

  • Magistral — the first reasoning model by Mistral AI
    Subscribe Sponsored by: Dynatrace — When agents enter the SDLC, observability becomes the enabler to …Subscribe Sponsored by: Dynatrace — When agents enter the SDLC, observability becomes the enabler to move from code generation to scalable engineering. Read the blog for a framework to get started 10…

    simonwillison.net

  • mistralai/Magistral-Small-2506 at main
    mistralai / Magistral-Small-2506 like 610 Follow Mistral AI_ 18.8k Safetensors 24 languages vllm mistral mistral-common …mistralai / Magistral-Small-2506 like 610 Follow Mistral AI_ 18.8k Safetensors 24 languages vllm mistral mistral-common arxiv: 2506.10910 License: apache-2.0 Model card Files Files and versions xet C…

    huggingface.co

  • Paper page - Magistral - Hugging Face
    this paper, we introduce Mistral’s first reasoning models: Magistral Small and Magistral Medium, based on …this paper, we introduce Mistral’s first reasoning models: Magistral Small and Magistral Medium, based on the Mistral Small 3 and Mistral Medium 3 models respectively, and outline our proposed RLVR f…

    huggingface.co

  • README.md · mistralai/Magistral-Small-2509 at main
    Mistral AI_ 18.8k Safetensors 24 languages vllm mistral3 mistral-common arxiv: 2506.10910 License: apache-2.0 Model card …Mistral AI_ 18.8k Safetensors 24 languages vllm mistral3 mistral-common arxiv: 2506.10910 License: apache-2.0 Model card Files Files and versions xet Community 9 Copy to bucket new main Magistral-Sma…

    huggingface.co

  • README.md · mistralai/Magistral-Small-2506 at main
    README.md · mistralai/Magistral-Small-2506 at main Hugging Face Models Datasets Spaces Community Docs Enterprise Pricing Log …README.md · mistralai/Magistral-Small-2506 at main Hugging Face Models Datasets Spaces Community Docs Enterprise Pricing Log In Sign Up mistralai / Magistral-Small-2506 like 599 Follow Mistral AI_ 12…

    huggingface.co

  • mistralai/Magistral-Small-2509 · Hugging Face
    Updates compared with Magistral Small 1.1 Key Features Benchmark Results Sampling parameters Basic Chat Template …Updates compared with Magistral Small 1.1 Key Features Benchmark Results Sampling parameters Basic Chat Template Usage Inference Fine-tuning vLLM (recommended) Transformers Magistral Small 1.2 Sectio…

    huggingface.co

  • Magistral - arXiv.org
    https://mistral.ai/news/mistral-large-2407 , 2024. MistralAI. Mistral medium 3. https://mistral.ai/fr/news/mistral-medium-3 , 2025. Michael Noukhovitch, Shengyi Huang, Sophie …https://mistral.ai/news/mistral-large-2407 , 2024. MistralAI. Mistral medium 3. https://mistral.ai/fr/news/mistral-medium-3 , 2025. Michael Noukhovitch, Shengyi Huang, Sophie Xhonneux, Arian Hosseini…

    arxiv.org

Also known as: magistral-medium-2506, magistral-medium-latest, mistral/magistral-medium, mistral/magistral-medium-latest, mistral/magistral-medium-latest@eu, mistralai/magistral-medium-2506, mistralai/magistral-medium-2506:thinking, vercel_ai_gateway/mistral/magistral-medium

Like this one. Tap or hover any dotted value, label or icon to see where it comes from and how it is measured. open details when tapped or hovered. How this data works